🔭 astrophysics

Serial Charge Transfer Efficiency in ACS/WFC

This study characterizes the serial charge transfer efficiency (CTE) of the ACS/WFC CCDs using dark frame hot pixel trails, quantifies its impact on photometry and astrometry, and implements a new pixel-based correction in CALACS to mitigate these losses for post-SM4 full-frame images.

Original authors: Jenna E. Ryon, Norman A. Grogin

Published 2026-02-04

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Jenna E. Ryon, Norman A. Grogin

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine the Hubble Space Telescope's "Wide Field Camera" (ACS/WFC) as a giant, incredibly sensitive digital camera. Like any digital camera, it captures light and turns it into an image. But this camera is old, and it has been orbiting Earth for decades, bombarded by cosmic radiation. This radiation has damaged the camera's internal sensors, creating tiny "traps" inside the pixels.

This report is about fixing a specific type of "blur" caused by these traps, known as Charge Transfer Efficiency (CTE).

Here is the breakdown of what the scientists found and fixed, using simple analogies:

1. The Problem: The "Muddy Walk"

When the camera takes a picture, it doesn't just snap a photo; it has to physically move the captured light (electrons) from the sensor to the computer to be read out. Think of this like a line of people passing buckets of water down a human chain to a fire hydrant.

The Damage: Over time, the "people" in the chain (the pixels) have developed potholes (charge traps). As the buckets of water pass through, some water gets stuck in the potholes.
The Result: By the time the water reaches the end, the bucket is lighter than it should be, and there are little puddles of water left behind in the potholes along the path. In the final image, this looks like a faint "tail" or trail behind bright stars, and the stars themselves look slightly dimmer and slightly shifted in position.

2. Two Directions, Two Problems

The camera moves the water in two directions:

Vertical (Parallel): Moving row by row. This is a long, slow walk. Because it takes a long time, a lot of water gets stuck in the potholes. This is the "big" problem scientists have been fixing for years.
Horizontal (Serial): Moving pixel by pixel across a row. This is a very fast sprint. Because it happens so quickly, fewer potholes have time to grab the water.

The Discovery: While the horizontal (serial) problem is much smaller than the vertical one, it's not zero. It's like a sprinter who usually runs perfectly but occasionally trips over a tiny pebble. For most photos, you wouldn't notice. But for scientists measuring the exact brightness or position of faint stars, that tiny trip matters.

3. How They Studied It

The scientists couldn't easily see these tiny trails in normal star photos because the stars are too complex. Instead, they used "Hot Pixels."

The Analogy: Imagine a hot pixel is a person in the line holding a massive, glowing bucket of water (a "delta function"). Because the bucket is so big and bright, if even a tiny drop gets stuck in a pothole, you can clearly see the trail left behind.
The Method: They looked at thousands of "dark" photos (pictures taken with the lens cap on) to find these hot pixels. They measured exactly how much water was left behind in the trail for different sizes of buckets. This allowed them to map out exactly where the potholes are and how sticky they are.

4. The Solution: A Digital "Sponge"

The team updated the software (called CALACS) that processes the telescope's images.

The Old Way: The software knew how to fix the vertical trails (the long walk).
The New Way: They added a new step to also fix the horizontal trails (the sprint).
How it Works: The software acts like a smart sponge. It looks at the image, calculates how much water was likely lost to the potholes based on the star's position and brightness, and then "pours" that missing water back into the star. It also shifts the star's position back to where it belongs.

5. What Changed?

The report claims that with this new fix:

Brightness: Stars are now measured to be about 0.005 to 0.02 magnitudes brighter (correcting for the lost water).
Position: The location of stars is shifted back by about 0.01 to 0.035 pixels.
Visuals: In the "dark" test images, the long, messy trails behind hot pixels have almost completely vanished. The stars look like sharp, single points again.

6. Who Gets the Fix?

When: This fix is applied to all full-frame images taken after May 2009 (after a major repair mission called SM4). Before that, the camera was so new that the horizontal trails were negligible.
Where: The corrected data is being made available in the public archive (MAST) for astronomers to use.

In Summary:
The Hubble camera has a slight "stutter" when moving data sideways. This report describes how scientists mapped out the exact nature of that stutter using "hot pixels" as test subjects and updated the camera's software to automatically smooth out the data. The result is sharper, more accurate images for scientists studying the universe.

Technical Summary: Serial Charge Transfer Efficiency in ACS/WFC

Problem Statement
The Advanced Camera for Surveys (ACS) Wide Field Channel (WFC) on the Hubble Space Telescope has long been characterized by Charge Transfer Efficiency (CTE) degradation in the parallel (y-direction) transfer due to radiation damage. While parallel CTE effects are well-studied and corrected, the serial (x-direction) CTE has historically been considered negligible due to the extremely short dwell time between serial transfers (22 µs) compared to parallel transfers (3212 µs). However, as the instrument ages, serial CTE degradation warrants dedicated investigation. Unaddressed serial CTE causes charge trapping and delayed release, creating trails behind sources that degrade photometry (flux loss) and astrometry (centroid shifts), particularly for faint sources on low backgrounds. Previous studies noted the effect but lacked a quantified, pixel-based correction model.

Methodology
The authors developed an empirical model for serial CTE using data from the CCD Daily Monitor calibration program.

Data Source: Long dark frames (1000.5s) stacked over calendar years, containing "hot pixels" (high dark current pixels) that act as delta-function charge sources.
Trail Analysis: The study isolated hot pixels with specific charge levels (e.g., 30,000 $e^-$ ) and analyzed the "serial trails" extending upstream (towards the amplifier) in the readout direction.
Statistical Processing: To detect faint trails, the authors combined data from entire calendar years and applied Kernel Density Estimation (KDE) to sigma-clipped pixel values, robustly determining the peak signal of the trail pixels and background levels.
Modeling:
- Trail Profile: A triple-exponential fit was applied to the trail shapes to model the probability of charge release as a function of distance from the source.
- Trap Density: The total deferred charge was used to estimate the cumulative number of charge traps encountered by charge packets of varying sizes ( $q$ ) across the 2048-pixel serial register.
- Iterative Refinement: Initial parameters were derived from 2022 data. The model was iteratively adjusted by running the correction on uncorrected data and manually tuning the trap density and trail profiles until residual trails in corrected dark frames approximated background noise.
Time Dependence: The model incorporates time-dependent scaling factors ( $t_0$ and $t_1$ ) to account for degradation over the mission lifetime, specifically distinguishing between pre-SM4 (Servicing Mission 4, May 2009) and post-SM4 eras.

Key Contributions

Quantification of Serial CTE: The study provides the first dedicated, quantitative characterization of serial CTE in ACS/WFC, determining specific trap density and trail profiles for each of the four detector quadrants (A, B, C, D).
Algorithm Integration: The pixel-based CTE correction algorithm in the CALACS software (specifically the ACSCTE step) was modified to include serial CTE correction. The correction is applied before parallel CTE correction to accurately reverse the readout process.
Reference File Update: The PCTETAB reference file was updated to include the new serial CTE parameters (trap density profiles, trail profiles, and time constants), making the correction available in the standard calibration pipeline.
Implementation Scope: The correction is implemented by default for full-frame WFC images obtained after SM4. It is not currently applied to subarray data or pre-SM4 data, where serial losses were deemed negligible.

Results

Trail Characteristics: Serial CTE trails are sharper and longer than parallel trails. The first pixel in a serial trail typically contains ~0.5–1.0% of the source signal, with subsequent pixels dropping off rapidly but remaining detectable for up to 100 pixels (and potentially further).
Photometric Impact: Serial CTE causes flux losses of approximately 0.005 to 0.02 magnitudes in stellar photometry, depending on stellar brightness and background level. The correction restores this lost signal.
Astrometric Impact: Serial CTE shifts stellar centroids by 0.01 to 0.035 pixels, with the magnitude of the shift increasing as the number of serial transfers increases (i.e., closer to the amplifier gap).
Quadrant Variations: Significant differences exist between quadrants. Quadrant D exhibits fainter trails, while Quadrant B shows slightly larger corrections (up to ~0.02 mag and ~0.035 pixels).
Time Evolution: Serial CTE losses were less than one-third of post-SM4 levels prior to SM4. Post-SM4 data shows a linear increase in residuals until 2022, followed by a potential overcorrection in 2023 data, suggesting non-linear evolution or temperature effects that required manual adjustment of time constants ( $t_0 = \text{MJD } 49000$ , $t_1 = \text{MJD } 60500$ ).

Significance and Claims
The paper asserts that while serial CTE effects are an order of magnitude smaller than parallel CTE effects, they are not negligible for high-precision photometry and astrometry. The implementation of this correction in CALACS v10.4.0 ensures that science data processed after the report's publication will be corrected for both parallel and serial CTE losses. The authors note that the correction significantly improves data quality, reducing flux loss and centroid shifts, though they acknowledge minor residual systematic biases in bright stars that may require future investigation. The study confirms that the WFC serial trail extends further than the current 100-pixel model, a topic reserved for future work.